Limb amputations can significantly negatively impact amputees' lives. Fortunately, prosthetic devices can partially compensate for loss of limb structure (bone, skin, etc.) and actuation (muscle). An ideal, albeit not yet developed, prosthetic device would include a separate motor for each lost muscle or degree of limb freedom, and each such motor would be driven by a discrete signal from the amputee's nervous system. Similarly, the ideal prosthetic device would include a sensor (touch, temperature, etc.) corresponding to each sense signal the amputated limb would otherwise have sent to the amputee's nervous system.
In a nervous system, efferent axons, otherwise known as motor or effector neurons, carry nerve impulses from a central nervous system to effectors, such as muscles and glands. On the other hand, afferent axons, otherwise known as sensory nerves or receptor neurons, carry nerve impulses from receptors or sense organs towards the central nervous system. Neural interfaces are, therefore, important for coupling efferent and afferent nerves to motors and sensors, respectively, in prosthetic devices.
Interfacing with efferent and afferent axons is difficult, at least in part due to their small sizes. In general, as illustrated in FIG. 1, a peripheral nerve 100 includes blood vessels 102 and several fascicles 104, each fascicle containing a bundle of axons 106. A typical mammalian fascicle is about 500 μm in diameter, but the diameter may vary depending on location and function of the nerve fiber. Fascicles in other animals may be other sizes.
In some prior art nerve interfaces, three electrodes are disposed longitudinally along a nerve or a fascicle. Two of the electrodes are used to establish an electrical reference voltage, and the third electrode provides an electrical measurement signal. However, the electrical measurement provides an integrated signal, i.e., a sum of signals from a plurality of axons in the nerve or fascicle. A fascicle contains a combination of efferent and afferent axons, and all the efferent axons typically do not control a single muscle. Thus, the integrated measurement signal is of limited value for selectively driving a motor of a prosthetic device.
Higher density neural interfaces, i.e., interfaces that provide an electrical signal from a small numbers of axons, ideally from one axon, would facilitate finer motor control in prosthetic devices than is achievable in the prior art. Similarly, higher density neural interfaces would facilitate more granular sensory feedback from prosthetic devices to central nervous systems.